Slashdot Mirror
PostgreSQL 8.1.4 Released to Plug Injection Hole
alurkar writes to tell us that PostgreSQL released version 8.1.4 today in order to combat a security flaw allowing a SQL injection attack. From the article: "The vulnerability affects PostgreSQL servers exposed to untrusted input, such as input coming from Web forms, in conjunction with multi-byte encodings like (Shift-JIS (SJIS), 8-bit Unicode Transformation Format (UTF-8), 16-bit Unicode Transformation Format (UTF-16), and BIG5. In particular, Berkus says that applications using 'ad-hoc methods to "escape" strings going into the database, such as regexes, or PHP3's addslashes() and magic_quotes' are particularly unsafe. 'Since these bypass database-specific code for safe handling of strings, many such applications will need to be re-written to become secure.'"
Most of the PHP apps I've ever had the (mis)pleasure to peruse make liberal use of this type of "escaping" rather than calling the provided "escape_string" functions. That never made any sense to me, but the practice appears to be quite common.
It especially bugs me because it's easier to Do Things Right. The DBI manpage for perl doesn't even mention the sloppy way that nearly everyone uses... but they do it anyway! In nearly every database application / script I look at, people do things like $dbh->execute("SELECT * FROM foo WHERE bar=$bar AND baz=$baz") after "escaping" $bar and $baz. No, no, no!
It's much easier to prepare a query handle and then execute it as needed:
$sth = $dbh->prepare("SELECT a,b,c FROM foo WHERE bar=? and baz=?")
$sth->execute($bar, $baz);
Not only is it more efficient (if you're going to use the same query twice), it's secure by default. Let the database programmers handle the Hard Stuff (parsing) so that you can concentrate on your application.
Speaking of which, is there a way to do this in PHP? I've never seen a PHP script that did anything like this (which is probably why bugtraq is 99% php SQL injection holes).
My other car is first.
Mismatches between different character encodings seem to have been responsible for vast swathes of security vulnerabilities over the past few years. The sooner everybody moves to programming languages and software that use Unicode natively, the more secure we will all be.
Unfortunately, the languages receiving the most attention for web development have abysmal Unicode support. PHP and Ruby haven't a clue, although the next version of PHP is supposed to be much better in this respect. Python developers can at least handle things fairly well, although it's still a bit of a pain in the neck.
This vulnerability is probably going to cause quite a few problems for people, as it's a client issue that will probably need whatever adapter you use to be updated. Here is the user guide to the vulnerability for PostgreSQL. psycopg should be fixed shortly.
Bogtha Bogtha Bogtha
PHP5's mysqli extension enables you to use prepared queries.
As long as you set the right multibyte string encoding in PHP via the multibyte string functions (specifically, the mb_internal_encoding function), the parser will catch the invalid multibyte sequence and fix it.
Move along, folks. No need to panic.
Most people probably aren't aware of it, but several years ago, I wrote a few short scripts for PHP 4 that specifically address this problem. Currently-supported database backends are MySQL and anything that DBX supports, but it wouldn't take much to adapt it to PostgreSQL.
It basically lets you write code like this:
It doesn't have the performance benefits that real prepared statements have, but I still find it handy for typical PHP4 database work.
The code is released under the MIT license, so feel free to use it.
For PHP, Zend_Db has a way of doing this which is very similar to the way you do it in Perl and Java. It's quite nice. There are other ways of doing this as well :) // get a Zend_Db_Adapter (basically a DB connection) // the sql with a placeholder for a parameter called 'id' :id'; // anyparameters are defined in the array. in this case, just 'id' // send the query
$db = getConnection();
$sql = 'select * from Foo where id =
$params = array('id' => $id);
$result = $db->query($sql, $params);
PostgreSQL ignored invalid UTF-8 sequences, meaning a ' character at the end of a incomplete sequence could cause only one ' to be seen by the parser when escaped.
See http://www.postgresql.org/docs/techdocs.50 for the details.
dtach - A tiny program that emulates the detach feat
PEAR::DB supports almost the exact same method.
p
$data = array('one',2);
(short)
$result = $db->query('select * from table where foo=? and bar=?',$data);
(prepare)
$stmt = $db->prepare('select * from table where foo=? and bar=?');
$result = $db->execute($stmt,$data);
Works with mysql, pgsql, mssql... etc etc. MDB2 is the new version of this library which uses much the same syntax. Uses database-specific escaping/quoting automatically.
http://pear.php.net/manual/en/package.database.ph
--onyx--
Yes, but properly escaping everything is at least as important. Whitelisting and blacklisting can't be used in a lot of situations (for example text fields) without causing problems. The easiest way to do that is to use an existing library that handles most of that for you. The more you have automated, the less room there is for human error. Unfortunatly, PHP coders tend to trive in reinventing the wheel.